Re: [m5-dev] Functional Interface in Ruby

I'm also confused by what you mean when you say "moving all controllers from being under system to being under system.ruby". Remember that the cycle has nothing to do with parent/child relationships, it only has to do with cases where a SimObject pointer is passed as a parameter to another SimObje

Re: [m5-dev] Functional Interface in Ruby

I'm not sure this is a great solution, since eventually it would be nice to get rid of RubySystem as a separate object and just use System. (There's really no non-historical reason to have both.) I still don't quite understand where the cycles are coming from; the outputs you sent after adding th

Re: [m5-dev] Functional Interface in Ruby

In order to solve the cycle problem, I am thinking of moving all controllers from being under system to being under system.ruby. I would add them to RubySystem, after that these controllers can be used to construct a list of cache memories that are present in the system. Similarly, I will also

Re: [m5-dev] Object Initialization using Python

Nate, Gabe, Thanks, I am able to make use of VectorParam. Nilay On Thu, 10 Mar 2011, Gabe Black wrote: Here's one: http://repo.m5sim.org/m5/file/77aa0f94e7f2/src/cpu/BaseCPU.py#l91 Gabe On 03/10/11 16:40, nathan binkert wrote: I'm traveling so I don't have Access to code. Just grep for Ve

Re: [m5-dev] Review Request: Standard switch fix for multi-core

Maybe it's malformed? Gabe On 03/10/11 17:55, Korey Sewell wrote: > For some reason, it looks like we can download this diff but not view > it in the web interface... > > On Thu, Mar 10, 2011 at 6:05 PM, Anthony Gutierrez > wrote: > > >

Re: [m5-dev] Review Request: Standard switch fix for multi-core

For some reason, it looks like we can download this diff but not view it in the web interface... On Thu, Mar 10, 2011 at 6:05 PM, Anthony Gutierrez wrote: > > --- > This is an automatically generated e-mail. To reply, visit: > http://reviews

Re: [m5-dev] Object Initialization using Python

Here's one: http://repo.m5sim.org/m5/file/77aa0f94e7f2/src/cpu/BaseCPU.py#l91 Gabe On 03/10/11 16:40, nathan binkert wrote: > I'm traveling so I don't have Access to code. Just grep for VectorParam. > > Nate > > On Thu, Mar 10, 2011 at 3:55 PM, nathan binkert wrote: >> I'm traveling so I don'

Re: [m5-dev] Object Initialization using Python

I'm traveling so I don't have Access to code. Just grep for VectorParam. Nate On Thu, Mar 10, 2011 at 3:55 PM, nathan binkert wrote: > I'm traveling so I don't have Access to code. Just grep for VectorParam > > On Mar 10, 2011, at 15:01, Nilay Vaish wrote: > >> Can you point out an example of

Re: [m5-dev] disabling FastAlloc for m5.debug

I haven't had a chance to really consider it carefully, but it seems reasonable to me. One danger might be if the fast allocator changes behavior in such a way that bugs that show up in m5.opt go away in m5.debug when it's turned off. I've run into sort of the opposite of that before I think with v

Re: [m5-dev] Review Request: Standard switch fix for multi-core

--- This is an automatically generated e-mail. To reply, visit: http://reviews.m5sim.org/r/560/ --- (Updated 2011-03-10 15:05:23.673150) Review request for Default, Ali Saidi, Gabe Bl

Re: [m5-dev] Object Initialization using Python

Can you point out an example of vector parameter? That should serve my purpose. Thanks Nilay On Thu, 10 Mar 2011, nathan binkert wrote: As I understand, we use Python objects to initialize C++ objects. Is it possible to pass a pointer to an array (dynamic sized) from Python to C++? What do

Re: [m5-dev] disabling FastAlloc for m5.debug

> - There's some code protected by '#if FAST_ALLOC_DEBUG' that is intended to > help debug memory leaks using FastAlloc.  I've used it once or twice, but I > kind of doubt that anyone else has, and I'm sure it's not as good as using > valgrind or probably several other malloc debug packages.  It se

Re: [m5-dev] Object Initialization using Python

> As I understand, we use Python objects to initialize C++ objects. Is it > possible to pass a pointer to an array (dynamic sized) from Python to C++? What do you mean a pointer? We do have VectorParams that allow us to pass an stl vector of any parameter type to a class. If you mean a SimObject

Re: [m5-dev] disabling FastAlloc for m5.debug

I didn't do any comprehensive testing, but averaging a few short runs of the ruby tester on m5.opt showed compiling with FastAlloc about 3% faster than not. Steve On Thu, Mar 10, 2011 at 2:11 PM, nathan binkert wrote: > Seems reasonable. One question. Do we know that FAST_ALLOC actually > is

[m5-dev] Object Initialization using Python

As I understand, we use Python objects to initialize C++ objects. Is it possible to pass a pointer to an array (dynamic sized) from Python to C++? Thanks Nilay ___ m5-dev mailing list m5-dev@m5sim.org http://m5sim.org/mailman/listinfo/m5-dev

Re: [m5-dev] disabling FastAlloc for m5.debug

Seems reasonable. One question. Do we know that FAST_ALLOC actually is faster than the built in allocator? It's pretty darned old and memory allocators are pretty modern now. Nate On Thu, Mar 10, 2011 at 8:23 AM, Steve Reinhardt wrote: > Brad pointed out that he habitually sets NO_FAST_ALLO

Re: [m5-dev] Review Request: SCons: Stop embedding the mercurial revision into the binary.

--- This is an automatically generated e-mail. To reply, visit: http://reviews.m5sim.org/r/557/#review939 --- Ship it! looks right to me. - Nathan On 2011-03-10 11:37:11, Gabe Blac

Re: [m5-dev] Review Request: Gems: Eliminate the now unused GEMS_ROOT scons variable.

--- This is an automatically generated e-mail. To reply, visit: http://reviews.m5sim.org/r/556/#review938 --- Ship it! - Nathan On 2011-03-10 11:04:07, Gabe Black wrote: > > --

Re: [m5-dev] Review Request: Standard switch fix for multi-core

Anthony, Something happened and you didn't upload your diff. Nate On Thu, Mar 10, 2011 at 1:17 PM, Anthony Gutierrez wrote: >This is an automatically generated e-mail. To reply, visit: > http://reviews.m5sim.org/r/560/ > Review request for Default, Ali Saidi, Gabe Black, Steve Reinhardt

Re: [m5-dev] Functional Interface in Ruby

We really need to turn these into DPRINTFs. Does anyone want to volunteer to implement DPRINTF in python? It should be trivial, but I'm not available to do it for at least a few days. Nate On Thu, Mar 10, 2011 at 1:52 PM, Nilay Vaish wrote: > I had originally put a print statement in getCCO

Re: [m5-dev] Functional Interface in Ruby

I made some changes to MESI_CMP_directory.py and those changes are reflected in the output. The l1 and l2 controllers are now being attached to ruby instead of system. Nilay On Thu, 10 Mar 2011, Nilay Vaish wrote: I had originally put a print statement in getCCObject(), so using the word 'Cr

Re: [m5-dev] Functional Interface in Ruby

I had originally put a print statement in getCCObject(), so using the word 'Creating' might have been mis-leading. For the previous output that I sent, I had removed that statement. What follows is patch and then the output -- diff --git a/src/python/m5/SimObject.py b/src/python/m5/SimObject.p

Re: [m5-dev] Review Request: Ruby: Get rid of the dead ruby tester.

Thanks for the explanation... I was pretty sure the ruby tester still lived on, I just didn't realize it had moved somewhere else. Steve On Thu, Mar 10, 2011 at 12:23 PM, Beckmann, Brad wrote: > Gabe, thanks for putting this patch out for review. I had forgotten that > this directory still exis

[m5-dev] Review Request: Standard switch fix for multi-core

--- This is an automatically generated e-mail. To reply, visit: http://reviews.m5sim.org/r/560/ --- Review request for Default, Ali Saidi, Gabe Black, Steve Reinhardt, and Nathan Binke

Re: [m5-dev] Review Request: Ruby: Get rid of the dead ruby tester.

Gabe, thanks for putting this patch out for review. I had forgotten that this directory still exists. I moved the code that I'm most familiar with out of this directory last year, but I didn't touch the Racey tester code because I wasn't sure what to do with it. I believe that code was writte

[m5-dev] Review Request: SCons: Stop embedding the mercurial revision into the binary.

--- This is an automatically generated e-mail. To reply, visit: http://reviews.m5sim.org/r/557/ --- Review request for Default, Ali Saidi, Gabe Black, Steve Reinhardt, and Nathan Binke

Re: [m5-dev] Review Request: Ruby: Get rid of the dead ruby tester.

I don't think it's dead, just sleeping... I'm not sure why it's not compilable right now (I thought it was usable), but I'd rather just fix that up than whack the code. We definitely need some input from Brad or the Wisconsin folks before making this change. Steve On Thu, Mar 10, 2011 at 11:03 A

[m5-dev] Review Request: Gems: Eliminate the now unused GEMS_ROOT scons variable.

--- This is an automatically generated e-mail. To reply, visit: http://reviews.m5sim.org/r/556/ --- Review request for Default, Ali Saidi, Gabe Black, Steve Reinhardt, and Nathan Binke

[m5-dev] Review Request: Ruby: Get rid of the dead ruby tester.

--- This is an automatically generated e-mail. To reply, visit: http://reviews.m5sim.org/r/555/ --- Review request for Default, Ali Saidi, Gabe Black, Steve Reinhardt, and Nathan Binke

[m5-dev] disabling FastAlloc for m5.debug

Brad pointed out that he habitually sets NO_FAST_ALLOC when compiling m5.debug because the way FastAlloc handles memory allocation tends to be tolerant of memory allocation errors (like reusing objects after they've been deleted), and he's run across other people's bugs simply by running with NO_FA

Re: [m5-dev] Captcha on wiki

Thanks! I hope this helps. Steve On Wed, Mar 9, 2011 at 10:56 PM, Ali Saidi wrote: > Everyone, > > I've just added math captcha to the wiki to prevent spam. Please let me > know if you run into any issues. > > Thanks, > Ali > > ___ > m5-dev mailing l

Re: [m5-dev] Functional Interface in Ruby

Is that it? It seems like there should be more output than from your previous example, not less... On Thu, Mar 10, 2011 at 5:36 AM, Nilay Vaish wrote: > Steve, here is the output after putting in the print statements. > > Creating root params > Creating root > Done creating root > Creating

Re: [m5-dev] Functional Interface in Ruby

On Wed, Mar 9, 2011 at 5:30 PM, Beckmann, Brad wrote: > I believe the L1DcacheMemory is created right after system because inside > each protocol file the first thing attached to the system is the l1 > controllers. That way the controllers get a more descriptive name than what > they are as relat

[m5-dev] Assertion in o3 CPU inst queue impl

Hi, We've looked into the following piece of code in src/cpu/o3/inst_queue_impl.hh (http://repo.m5sim.org/m5/diff/7bb1223f9645/src/cpu/o3/inst_queue_impl.hh), it looks like the old assertion "assert(ns_inst_it != nonSpecInsts.end" should be removed when both cases are covered in the patch diff

Re: [m5-dev] Functional Interface in Ruby

Steve, here is the output after putting in the print statements. Creating root params Creating root Done creating root Creating system params Creating system Done creating system Creating system.l1_cntrl0 params Nilay On Wed, 9 Mar 2011, Steve Reinhardt wrote: It seems odd that it t

[m5-dev] Cron /z/m5/regression/do-regression quick

* build/ALPHA_SE/tests/fast/quick/00.hello/alpha/linux/inorder-timing passed. * build/ALPHA_SE/tests/fast/quick/00.hello/alpha/linux/simple-atomic passed. * build/ALPHA_SE/tests/fast/quick/00.hello/alpha/linux/simple-timing passed. * build/ALPHA_SE/tests/fast/quick/00.hello/alpha/l